Micro-Expression Recognition Based on Dual-Stream Motion-Anchored Cross-Fusion Network
Sensors · 06.06.2026 · 26 (12) · 3628
DOI: 10.3390/s26123628 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 06.06.2026. Verfasst von Junxian Li und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Information Engineering, Yangzhou Polytechnic University, Yangzhou 225009, China. Hauptthemen laut MEDLINE: Facial Expression, Pattern Recognition, Automated. Die Arbeit nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Jiangsu Provincial Science and Technology Deputy General Manager Program, Key Research Project of the "14th Five-Year Plan" Scientific Research Program of Jiangsu Open University (The City Vocational College of Jiangsu).
Abstract (Originalsprache)
Micro-expression recognition (MER) remains a formidable challenge in affective computing due to the subtle, localized, and fleeting nature of facial muscle actuations. Conventional spatial-temporal networks are easily overwhelmed by static facial topologies, leading to feature representations that are heavily biased toward identity-specific noise. To address this, we propose the Motion-Anchored Cross-Modal Fusion Network (MACFN), a novel dual-stream ViT architecture that explicitly decouples and synergizes spatial appearance and optical flow dynamics. Specifically, we introduce a motion-anchored spatial attention module, which translates latent motion features into a sparse spatial probability mask. It acts as an enhancement gate, forcing the texture stream to bypass static backgrounds and attend to genuine ME-related regions. Furthermore, we design a cross-modal bilinear fusion module to capture the second-order interactions across modalities, mapping the coupled features into a discriminative semantic manifold. Extensive experiments conducted on the CASME II, SAMM, and SMIC databases under the rigorous leave-one-subject-out composite database evaluation protocol demonstrate that MACFN is effective and achieves competitive performance compared to several recent methods.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Information Engineering, Yangzhou Polytechnic University, Yangzhou 225009, China.
- Suzhou Institute of Technology, Jiangsu University of Science and Technology, Suzhou 215699, China.
- School of Computer, Jiangsu University of Science and Technology, Zhenjiang 212003, China.
- School of Computing and Artificial Intelligence, Southwest Jiaotong University, Chengdu 611756, China.
- JingJiang College, Jiangsu University, Zhenjiang 212013, China.
ORCID-Profile: Junxian Li, Tian Li, Shucheng Huang, Gang Wang, Mingxing Li
Förderung
- National Natural Science Foundation of China
- Jiangsu Provincial Science and Technology Deputy General Manager Program
- Key Research Project of the "14th Five-Year Plan" Scientific Research Program of Jiangsu Open University (The City Vocational College of Jiangsu)
- Key Research Project of the "14th Five-Year Plan" Scientific Research Program of Jiangsu Open University
Zitieren
APA
Li, J., Li, T., Huang, S., Wang, G., & Li, M. (2026). Micro-Expression Recognition Based on Dual-Stream Motion-Anchored Cross-Fusion Network. Sensors, 26(12), 3628. https://doi.org/10.3390/s26123628
Vancouver
Li J, Li T, Huang S, Wang G, Li M. Micro-Expression Recognition Based on Dual-Stream Motion-Anchored Cross-Fusion Network. Sensors. 2026;26(12):3628. doi:10.3390/s26123628
BibTeX
@article{Li2026,
author = {Li, Junxian and Li, Tian and Huang, Shucheng and Wang, Gang and Li, Mingxing},
title = {Micro-Expression Recognition Based on Dual-Stream Motion-Anchored Cross-Fusion Network},
journal = {Sensors},
year = {2026},
volume = {26},
number = {12},
pages = {3628},
doi = {10.3390/s26123628}
}
EUVIMED